currently, the model outputs the rain, and so need a subtraction in filter c code to get the final derain result.
I've sent a PR to update the model file and accepted, see at https://github.com/XueweiMeng/derain_filter/pull/3 Signed-off-by: Guo, Yejun <yejun....@intel.com> --- doc/filters.texi | 7 +++++-- libavfilter/vf_derain.c | 6 +----- 2 files changed, 6 insertions(+), 7 deletions(-) diff --git a/doc/filters.texi b/doc/filters.texi index 328e984..3dc7715 100644 --- a/doc/filters.texi +++ b/doc/filters.texi @@ -8878,6 +8878,7 @@ delogo=x=0:y=0:w=100:h=77:band=10 @end itemize +@anchor{derain} @section derain Remove the rain in the input image/video by applying the derain methods based on @@ -8932,6 +8933,8 @@ Note that different backends use different file formats. TensorFlow and native backend can load files for only its format. @end table +It can also be finished with @ref{dnn_processing} filter. + @section deshake Attempt to fix small changes in horizontal and/or vertical shift. This @@ -9201,9 +9204,9 @@ Set the output name of the dnn network. @itemize @item -Halve the red channle of the frame with format rgb24: +Remove rain in rgb24 frame with can.pb (see @ref{derain} filter): @example -ffmpeg -i input.jpg -vf format=rgb24,dnn_processing=model=halve_first_channel.model:input=dnn_in:output=dnn_out:dnn_backend=native out.native.png +./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg @end example @item diff --git a/libavfilter/vf_derain.c b/libavfilter/vf_derain.c index 89f9d5a..7432260 100644 --- a/libavfilter/vf_derain.c +++ b/libavfilter/vf_derain.c @@ -100,7 +100,6 @@ static int filter_frame(AVFilterLink *inlink, AVFrame *in) AVFilterLink *outlink = ctx->outputs[0]; DRContext *dr_context = ctx->priv; DNNReturnType dnn_result; - int pad_size; AVFrame *out = ff_get_video_buffer(outlink, outlink->w, outlink->h); if (!out) { @@ -129,15 +128,12 @@ static int filter_frame(AVFilterLink *inlink, AVFrame *in) out->width = dr_context->output.width; outlink->h = dr_context->output.height; outlink->w = dr_context->output.width; - pad_size = (in->height - out->height) >> 1; for (int i = 0; i < out->height; i++){ for(int j = 0; j < out->width * 3; j++){ int k = i * out->linesize[0] + j; int t = i * out->width * 3 + j; - - int t_in = (i + pad_size) * in->width * 3 + j + pad_size * 3; - out->data[0][k] = CLIP((int)((((float *)dr_context->input.data)[t_in] - ((float *)dr_context->output.data)[t]) * 255), 0, 255); + out->data[0][k] = CLIP((int)((((float *)dr_context->output.data)[t]) * 255), 0, 255); } } -- 2.7.4 _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
